Dichromatic Illumination?

[Nubee-70]

I have been thinking about how to blend two coloured LEDs to make a composite colour, such as red and blue to make purple.
If I mixed any two primary coloured LEDs (red, yellow, green, blue, and white) then I ought to be able to make almost any colour I want.
To do this, I figure that each LED ought to be able to change its own brightness, and duty-cycle intensity, via the use of a Pulse Width Modulation (PWM) circuit.

However, there are a few more problems to do this cheaply. Because of this I will be restricted to only use ordinary bright 5mm LEDs.
They will need to have some sort of fixed focal length where each LED zeros-in (overlapping) onto the specimen.
Due to less than ideal degree of brightness, I am assuming that the best time to use this illuminator would be less lit conditions, such as in the evening/night.

So, to make it a bit brighter, I think two LEDs per colour may suffice. That is, 10 LEDs all up, in a circular area with a diameter of about 20 mm, and perhaps about 23-25mm with shielding.

About the electronic circuit (PWM). Usually this circuit is used for varing the on-off time of a LED, acting as a dimmer circuit.
But in this case there are two LEDs involved. As one LED (red) stays on longer, within one duty cycle, the other LED (blue) stays on less. Making an impression that the illumination is more red than blue _ such as magenta.
The duty-cycle needs to be above 2Khz so that our nervous system, so the eyes, cannot sense the pulses.

So, this PWM needs to split the pulse into two, and invert one of them. That is, instead of on-off, it will be on-off, off-on.
If the duty cycle was 75% on and 25% off for the red LED, then the Blue LED will be 75% off and 25% on.
However, I think the change in duty-cycle will also effect the LED's brightness too. If this is the case, I will need to tie the selected LEDs to a resistor type dimmer, as part of its current limiting resistor. To act as a variable 'bias'.

This project is part of 'My Hobby Scope Set-up (WIP)' @ viewtopic.php?f=12&p=140580#p140575

A block diagram may look something like this...


So I will be designing this circuit and testing it on a breadboard. I will keep you informed.

[Hobbyst46]

It might help to know which type of microscope illumination is wished. Both brightfield and fluorescence, at least, benefit from a single die LED source, rather than a multi-LED arrangement. Besides, I wonder about the wavelength combination... For visual observation, green or blue or amber light are often appropriate, as alternatives to bright white. And these are easily achieved with simple filters and white light. For fluorescence, the combination of LED lights as described is unsuitable anyway (just my opinion).

[Scarodactyl]

Why use PWM rather than a constant current driver?

[MichaelG.]

It may be worth exploring these: /https://www.adafruit.com/category/168url
.
Ada deserves a lot of credit for her development work, but I should mention that even-cheaper ‘clones’ are to be found on ebay.

MichaelG.

[Nubee-70]

It might help to know which type of microscope illumination is wished. Both brightfield and fluorescence, at least, benefit from a single die LED source, rather than a multi-LED arrangement. Besides, I wonder about the wavelength combination... For visual observation, green or blue or amber light are often appropriate, as alternatives to bright white. And these are easily achieved with simple filters and white light. For fluorescence, the combination of LED lights as described is unsuitable anyway (just my opinion).

I can only afford the microscope I have _ a brightfield type. I wish to explore using a various choice of coloured illumination.
I cannot afford what the microscope industry offers up as fluorescence (as far as I know). Besides, I am thinking about adding two more specified light spectrum ranges of IR and UV.

Thanks for sharing your opinions on this, it has given me something else to think about.

[Nubee-70]

Why use PWM rather than a constant current driver?

So I can alter the degree of chromicity between the two coloured LEDs; offering a variable colour selection between the two without altering the degree of required illumination.

[Nubee-70]

It may be worth exploring these: /https://www.adafruit.com/category/168url
.
Ada deserves a lot of credit for her development work, but I should mention that even-cheaper ‘clones’ are to be found on ebay.

MichaelG.

Sorry, but the link does not go to what you want to show me.

[MichaelG.]

I will try again
MichaelG.

https://www.adafruit.com/category/168

[PeteM]

There will be some issues with properly collimating two LEDS - and then the results might not be all that interesting. Other users have swapped in IR and UV sources - but each was positioned centrally to match the old tungsten filament position.

Some other possibilities you might want to try:

1) Illuminate with a wide spectrum and then change colors in post-processing.

2) Use polarization and a wave plate with birefringent specimens.

3) DIY Rheinberg illumination for a wide variety of optical staining effects: https://www.olympus-lifescience.com/en/ ... rheinberg/

My recollection is that Saul, perhaps another source, sells affordable sets of Rheinberg filters to fit various condensers. A web search will reveal multiple DIY approaches.

[Nubee-70]

There will be some issues with properly collimating two LEDS - and then the results might not be all that interesting. Other users have swapped in IR and UV sources - but each was positioned centrally to match the old tungsten filament position.

Some other possibilities you might want to try:

1) Illuminate with a wide spectrum and then change colors in post-processing.

2) Use polarization and a wave plate with birefringent specimens.

3) DIY Rheinberg illumination for a wide variety of optical staining effects: https://www.olympus-lifescience.com/en/ ... rheinberg/

My recollection is that Saul, perhaps another source, sells affordable sets of Rheinberg filters to fit various condensers. A web search will reveal multiple DIY approaches.

Cool. Thank you PeteM.
Very interesting.

[Nubee-70]

I will try again
MichaelG.

https://www.adafruit.com/category/168

Cool
Thanks for the link.
I will be buying a few LEDs from them to play with. But at the moment need to see my idea through.
Perhaps I will be ending up making various LED applications as I gain more knowledge and resources.

[Scarodactyl]

Why use PWM rather than a constant current driver?

So I can alter the degree of chromicity between the two coloured LEDs; offering a variable colour selection between the two without altering the degree of required illumination.

I'd think you could do the same with constant current since they can vary the current they provide if you get the right controller, and without any of the issues cameras can get with flickering leds.

[Nubee-70]

Why use PWM rather than a constant current driver?

So I can alter the degree of chromicity between the two coloured LEDs; offering a variable colour selection between the two without altering the degree of required illumination.

I'd think you could do the same with constant current since they can vary the current they provide if you get the right controller, and without any of the issues cameras can get with flickering leds.

Thank you Scarodactyl for the information. I will check it out.

[Nubee-70]

After three attempts I think I found a working circuit, but I will soon find out after a few parts are posted to me.
Then I will put them onto the breadboard to see if the circuit actually works. And after fine tuning it, I will finally know the component values.

I already have most of the parts, but not all the LEDs, the 4001 NOR gates IC, and some of the hardware (switches and knobs).



As you can see from the schematic, I have opted to use the invisible spectrum LEDs (IR and UV) separate from the visible colour LEDs. So, this means that I will be having two types of illumination probes.
Switching ON the invisible LEDs will initiate a warning flashing light, on the console, to remind me to use eye (IR & UV) protection glasses.

[GerryR]

It is interesting that I was thinking about a similar project. My solution was using a 555 astable at ~100hz and triggering four more 555s set up as one-shots with a variable period to regulate the intensity of the four LEDs, R, G, B, Y (poor mans PWM). Because they are all triggered at the same time, they are all synched together. Each LED intensity is separately controlled, so "mixing" for different colors is straight forward. I was thinking about funneling the LED outputs into a fiber optic cable for "pointing" the light to the proper location. "A little knowledge in the hands of a fool is dangerous!" I'm very dangerous! This might trigger some other ideas for you. (I like your use of IR and UV LEDs; never gave them a thought.)

[Nubee-70]

It is interesting that I was thinking about a similar project. My solution was using a 555 astable at ~100hz and triggering four more 555s set up as one-shots with a variable period to regulate the intensity of the four LEDs, R, G, B, Y (poor mans PWM). Because they are all triggered at the same time, they are all synched together. Each LED intensity is separately controlled, so "mixing" for different colors is straight forward. I was thinking about funneling the LED outputs into a fiber optic cable for "pointing" the light to the proper location. "A little knowledge in the hands of a fool is dangerous!" I'm very dangerous! This might trigger some other ideas for you. (I like your use of IR and UV LEDs; never gave them a thought.)

Thanks GerryR for your shae about similar ideas. Interesting.

Currently, with this project, I will be using two flexible goosenecks with multi-pin socket in its end. The probe end, of various designs, can then be selected and plugged in to it.

RE: Optic fibres. I used to scratch build Sci-Fi models as a hobby, and found that if you want to us a single optic fibre (depending on its diameter) it is best to drill into the LED lens and insert the fibre (glued) and then wrap the LED in foil to minimize LED light leakage.

[GerryR]

Thanks GerryR for your shae about similar ideas. Interesting.

Currently, with this project, I will be using two flexible goosenecks with multi-pin socket in its end. The probe end, of various designs, can then be selected and plugged in to it.

RE: Optic fibres. I used to scratch build Sci-Fi models as a hobby, and found that if you want to us a single optic fibre (depending on its diameter) it is best to drill into the LED lens and insert the fibre (glued) and then wrap the LED in foil to minimize LED light leakage.

Looking forward to seeing your results!

[Nubee-70]

Thanks GerryR for your shae about similar ideas. Interesting.

Currently, with this project, I will be using two flexible goosenecks with multi-pin socket in its end. The probe end, of various designs, can then be selected and plugged in to it.

RE: Optic fibres. I used to scratch build Sci-Fi models as a hobby, and found that if you want to us a single optic fibre (depending on its diameter) it is best to drill into the LED lens and insert the fibre (glued) and then wrap the LED in foil to minimize LED light leakage.

Looking forward to seeing your results!

Same here. It just might become what we called in the RAAF a 'WAFWOTAM' _ 'what a ****** waste of time and money'.

[Nubee-70]

The schematic diagram has been updated to version 1.4.
Two sockets and plugs are shown, both are six pin.
The IR and UV probe can be operated/selected in the following ways: IR1. IR1+IR2. UV1. UV1+UV2. IR1+UV1. IR1+IR2+UV1+UV2.

[GerryR]

Thanks GerryR for your shae about similar ideas. Interesting.

Currently, with this project, I will be using two flexible goosenecks with multi-pin socket in its end. The probe end, of various designs, can then be selected and plugged in to it.

RE: Optic fibres. I used to scratch build Sci-Fi models as a hobby, and found that if you want to us a single optic fibre (depending on its diameter) it is best to drill into the LED lens and insert the fibre (glued) and then wrap the LED in foil to minimize LED light leakage.

Looking forward to seeing your results!

Same here. It just might become what we called in the RAAF a 'WAFWOTAM' _ 'what a ****** waste of time and money'.

There is ALWAYS something to be learned!!

[Nubee-70]

While waiting for electronic parts, I went ahead with fashioning a clear acrylic rod (20mm dia) to a point of about 5mm dia.
At the moment, it is just a rough job to see who it would work as a means to funnel the LED light to a small stage light.

I used a hacksaw to make a rectangular frustum with a point of about 5mm square, then started filing the edges into some sort of a conical frustum.

To see how the light may channel through the point, I used some masking tape to dampen the escaping light.
Placed a LED torch at wide end and see how it came out the pointed end. Personally, I am very pleased with the effect.



I plan to use aluminium foil to cover the light head; it may make a slightly brighter beam, as well as blocking the light leakage.

[jorymil]

I believe there's a flaw in your reasoning here. Combining light from two or three differently colored LEDs does _not_ generate light of any color. You can get light that appears to your eyes as any color you like, but the actual physical characteristics of that light are limited to the 20 nm or so bands of each LED. Say you want to generate violet light. You can tune red and blue LEDs such that light will appear violet to your eyes, but the actual light will diffract and refract as ordinary red and blue light. So depending on what you're looking to do (say you have a molecule that absorbs strongly at a particular wavelength), you may not get what you're expecting.

[MichaelG.]

… I plan to use aluminium foil to cover the light head; it may make a slightly brighter beam, as well as blocking the light leakage.

An impressive start !!

In an ideal world, however, you would polish those sawn faces and get total internal reflection

[ big trade-off in the realm of reward-for-effort though ]

MichaelG.

[Nubee-70]

I believe there's a flaw in your reasoning here. Combining light from two or three differently colored LEDs does _not_ generate light of any color. You can get light that appears to your eyes as any color you like, but the actual physical characteristics of that light are limited to the 20 nm or so bands of each LED. Say you want to generate violet light. You can tune red and blue LEDs such that light will appear violet to your eyes, but the actual light will diffract and refract as ordinary red and blue light. So depending on what you're looking to do (say you have a molecule that absorbs strongly at a particular wavelength), you may not get what you're expecting.

Thanks for sharing your knowledge. I will remember what you wrote when I wonder why I don't get the colour response I was expecting.

[Nubee-70]

...
In an ideal world, however, you would polish those sawn faces and get total internal reflection
[ big trade-off in the realm of reward-for-effort though ]
MichaelG.

Thanks.
I will clean it up to the point where I know that further polishing becomes inefficiently effective.

[GerryR]

I believe there's a flaw in your reasoning here. Combining light from two or three differently colored LEDs does _not_ generate light of any color. You can get light that appears to your eyes as any color you like, but the actual physical characteristics of that light are limited to the 20 nm or so bands of each LED. Say you want to generate violet light. You can tune red and blue LEDs such that light will appear violet to your eyes, but the actual light will diffract and refract as ordinary red and blue light. So depending on what you're looking to do (say you have a molecule that absorbs strongly at a particular wavelength), you may not get what you're expecting.

Very interesting. Can you explain further. In some light microscopy, I believe you are seeing reflective light (Incident lighting). So, in this case, if the specimen reflects, say, the red and the blue, will you not see violet? If you were to suppliment your standard lighting system with a variable color system, I don't see how you can do anything but enhance your view. The absorbtion process of a through (standard) lighting system I can understand, but it would still tell you something. For example, if you are sourcing red and blue and only see blue through the eyepiece, it would tell you that "something" is absorbing the the red. I don't profess to know the significance of such a scenario, but it is indicating something.

[MichaelG.]

The subject gets a bit heavy, Gerry … but this is a good introduction:

https://www.animations.physics.unsw.edu ... mixing.htm

The visual perception of a particular colour does not necessarily mean that colour’s actual wavelength is present.

MichaelG.

[GerryR]

The subject gets a bit heavy, Gerry … but this is a good introduction:

https://www.animations.physics.unsw.edu ... mixing.htm

The visual perception of a particular colour does not necessarily mean that colour’s actual wavelength is present.

MichaelG.

Thanks Michael. Most of that I was familiar with. I think, still, that the OPs project will still offer some interesting results, especially in a reflective light scenario; absorbtion-wise he will be limited to the LED's colors.

[Nubee-70]

The subject gets a bit heavy, Gerry … but this is a good introduction:
https://www.animations.physics.unsw.edu ... mixing.htm
The visual perception of a particular colour does not necessarily mean that colour’s actual wavelength is present.
MichaelG.

Thank you Michael for sharing the link to light spectrum colour mixing.

[MichaelG.]

… I think, still, that the OPs project will still offer some interesting results …

Most definitely … I’m looking forward to seeing them

MichaelG.